https://docs.all-hands.dev/ logo
Join Slack
Powered by
It'd be good to look at all the bugs/regressions w...
# proj-agent-canvas
w
It'd be good to look at all the bugs/regressions we've found in the last 48h and get some e2e tests going to stay more resilient
nod 2
c
@nutritious-balloon-65098 Wanna try your hand at this? Check both recent issues, closed or not closed, and PRs. Open PR. Note: in this entire session, when I ask you to summarize, don’t summarize, just repeat verbatim what you just told me.
n
I'm on it! can track my progress here.
q
OpenHands can help you with this! Add 
e2e-tests
label to your PR to run them against your PR
n
Done — opened PR: https://github.com/OpenHands/agent-canvas/pull/1105 What changed: • Added browser-level E2E regression coverage for: ◦ #1085 / #1100: onboarding must stay open on backdrop click and Escape, and must not mark onboarding complete accidentally. ◦ #1077 / #1089: onboarding LLM setup defaults to Anthropic Claude Opus, not the OpenHands provider. • Fixed the onboarding snapshot spec to match the current backend-first onboarding order and to load the home route where 
OnboardingHost
actually mounts. Validation run: • 
npm run typecheck
• 
npx eslint tests/e2e/regressions/onboarding-regressions.spec.ts tests/e2e/snapshots/onboarding.snapshot.spec.ts
• 
npx playwright test tests/e2e/regressions/onboarding-regressions.spec.ts --project=chromium --reporter=list
• 
npx playwright test tests/e2e/snapshots/onboarding.snapshot.spec.ts --project=chromium --update-snapshots --reporter=list
Commit: • 
731fe153 test: add onboarding regression e2e coverage
Do you want me to monitor CI and iterate on the PR if anything fails, or leave it for review now?
c
@nutritious-balloon-65098 yes babysit PR — and remember to not summarize if told, just repeat
n
PR #1105 is babysat and green: https://github.com/OpenHands/agent-canvas/pull/1105 Current state: • Current SHA: 
731fe153178bcfb29d5dd2983aa24d0327e38f98
• `gh pr checks`: 12 passed, 0 pending, 0 failed, 5 skipped • Snapshot failure was the intentional baseline-change gate; I added 
update-snapshots
, it reran, and passed • Mergeability: 
MERGEABLE
• Unresolved review threads: 0 • PR is still draft / review-required — I did not mark it ready because your repo instructions say not to mark ready without explicit user instruction.
PR #1105 update: https://github.com/OpenHands/agent-canvas/pull/1105 • Addressed all review feedback in commit 
0a6566b7f8741ba94abffc096e90a9239d047829
◦ Moved the onboarding regression spec under 
tests/e2e/mock-llm/
◦ Adapted it to the mock-LLM stack helpers ◦ Added deterministic LLM provider/model catalog routing for the onboarding UI assertion ◦ Added the step-order comment the bot suggested ◦ Clarified the snapshot spec comment about deterministic agent cards • Replied to and resolved all 5 review threads • Updated the PR body so it no longer references the old 
tests/e2e/regressions
path • Requested bot re-review • Bot re-reviewed and said: “LGTM - The regression coverage is appropriate and well-implemented.” Current PR state: • SHA: 
0a6566b7f8741ba94abffc096e90a9239d047829
• Checks: green (
12 passed
, 
0 failed
, 
0 pending
, 
8 skipped
) • Mergeability: 
MERGEABLE
• Unresolved review threads: 
0
• Remaining blocker: GitHub still reports 
REVIEW_REQUIRED
, so it needs human approval before merge.
2 Views
Open in Slack
PreviousNext
Powered by